Oh, I did not say I lost due to "objective" reasons, like technical features, stability and so on. I lost due to a very strong belief of the client that said "I want a standard product from a known company with a strong brand name". I was not able to successfully argue Perl against that belief.

For the technical perspective: I have worked with VisualAge for Java and WebSphere, and my conclusion is that this combination is a sure productivity killer (we spent days just configuring WebSphere), and unstable on top of that (deployment of EJBs is a true nightmare, with version incompatibilities and everything).

For more on the WebSphere technical side of this, see this collection of reviews.

VisualAge for Java is also highly unusual in its IDE approach, at least for those not having a Smalltalk background, and will guarantee a lot of confusion for people coming from a "normal" IDE background. (VisualAge takes the approach that you browse a class hierarchy, modifying attributes of classes and inventing new ones along the way. So, e.g., you will usually work on the source code of a single method only, with the rest of the class not shown in the editor window. To see code in another method, you have to change to a new window (or "frame"), forcing you to decide on a commit/cancel for your changes. But who cares anyway.)

Christian Lemburg
Brainbench MVP for Perl
http://www.brainbench.com